At the season level, Pittsburgh's 2025 pass defense looked like a near-photocopy of its 2024 unit by the number that gets quoted most: the Steelers allowed +0.030 pass EPA/play in 2024 and +0.028 in 2025 — 16th in the NFL both years. Same middle-of-the-pack pass defense, right?

Not quite. Underneath that flat top-line, one segment of the field completely caved in: the deep third. And it happened while the pass rush actually got better.

Table of Contents

• The top-line disguise
• The deep ball: a segment-specific collapse
• Everywhere else held or improved
• The pass rush wasn't the problem
• What it means

The top-line disguise

Here are the two Steelers pass defenses side by side at the season level:

Pittsburgh's sack rate climbed meaningfully — from 20th in the league to 12th — while pass EPA/play allowed barely moved, landing in the same 16th-place slot both years. If you stopped here, you'd say the pass rush got a bit better and the overall pass defense held steady.

The deep ball: a segment-specific collapse

Now bucket every pass thrown against Pittsburgh by air yards. Throws of 20+ air yards — the deep third — are where the entire story lives.

On roughly the same volume (54 → 49 attempts), the EPA allowed per deep throw more than quadrupled. Completion percentage went from 31.5% — elite — to 49.0%. Deep passing TDs allowed went from 2 to 8. And the takeaway edge that papered over any deep-ball leaks in 2024 (5 INTs) collapsed to a single pick.

The league-wide context sharpens the picture. In 2025, Pittsburgh's +0.979 EPA/play allowed on deep throws was the third-worst mark in the NFL, trailing only Minnesota (+1.08) and Dallas (+1.02). That same figure was a modest +0.231 in 2024 — nowhere near the bottom of the league by comparison.

Lamar Jackson went 3-of-3 for 140 yards and 2 deep touchdowns against Pittsburgh in Week 18 alone (+13.6 EPA on those three throws). Jared Goff hit 3-of-4 for 62 yards and a score in Week 16 (+5.31 EPA). Carson Wentz threw for 110 deep yards on 2 completions in Week 4 (+5.02 EPA). Shedeur Sanders (+5.9 EPA, Week 17), Caleb Williams (+4.0 EPA, Week 12), Joe Flacco (+4.64 EPA, Week 7) and Sam Darnold (+3.26 EPA, Week 2) each topped +3 EPA on a single deep shot against the Steelers.

Everywhere else held or improved

If this were a scheme-wide breakdown, every depth bucket would look worse. It doesn't:

Behind the line of scrimmage, Pittsburgh got better and quadrupled its takeaways in that zone (1 → 4 INTs). The short game got better too. The intermediate zone drifted slightly worse but the interception total (6 → 4) barely moved. Only the deep bucket cratered — and only the deep bucket lost its INT edge.

This isn't a coverage philosophy that stopped working across the board. It's one specific failure mode: contested deep throws are being completed instead of picked off.

The pass rush wasn't the problem

The standard explanation for a suddenly-porous deep defense is that the rush stopped getting home and quarterbacks had time to let routes develop. That's not what happened here.

Pittsburgh took down the quarterback on 7.21% of dropbacks in 2025, up from 6.36% — a jump from 20th to 12th in the NFL sack-rate rankings — while its QB-hit rate stayed essentially flat (16.38% to 16.22%). The Steelers were getting to the passer more often, not less, in the same season they gave up eight deep touchdowns.

What it means

The usual story is that pressure and coverage move together — you either affect the throw or you don't. Pittsburgh's 2025 pass defense complicates that assumption. The rush got home more often and climbed the league sack-rate rankings, the underneath coverage held or improved, and the deep third still doubled-plus in EPA allowed while losing 4 of its 5 interceptions from a year earlier.

Two seasons is a small sample and 49 deep attempts is a smaller one. Contested-catch outcomes swing on a handful of 50/50 balls, and going 5-for-5 on those in 2024 vs. 1-for-5 in 2025 could plausibly be partially noise. But the yards-per-attempt jump (11.06 → 16.33) and the 2 → 8 TD swing suggest something more than variance — receivers weren't just winning contested balls, they were running loose.